By Steven Goddard
I recently had the opportunity to attend a meeting of some top weather modelers. Weather models differ from climate models in that they have to work and are verified every hour of every day around the planet. If a weather model is broken, it becomes obvious immediately. By contrast, climate modelers have the advantage that they will be long since retired when their predictions don’t come to pass.
Weather and climate models are at the core very similar, but climate models also consider additional parameters that vary over time, like atmospheric composition. Climate models iterate over very long time periods, and thus compound error. Weather modelers understand that 72 hours is about the limit which they can claim accuracy. Climate modelers on the other hand are happy to run simulations for decades (because they know that they will be retired and no one will remember what they said) and because it provides an excuse to sink money into really cool HPC (High Performance Computing) clusters.
But enough gossip. I learned a few very interesting things at this meeting.
1. Weather modelers consider the realm of climate calculation to be “months to seasons.” Not the 30 year minimum we hear quoted all the time by AGW groupies. That is why NOAA’s “Climate Prediction Center” generates their seasonal forecasts, rather than the National Weather Service.
2. The two most important boundary conditions (inputs) to seasonal forecasts are sea surface temperatures and soil moisture. No one has shown any skill at modeling either of those, so no surprise that The Met Office Seasonal forecasts were consistently wrong.
For example, just a few months ago the odds of La Niña were considered very low. Compare the December forecast with the May version. How quickly things change!
SST modeling capabilities are very limited, and as a result seasonal weather forecasts (climate) are little more than academic exercises.
Oh and by the way, Colorado will be exactly 8.72 degrees warmer in 100 years. But they can’t tell you what the temperature will be next week.
“If I don’t understand it, it must be simple.”
– Dilbert Principle
In the top picture, which boxer is weather and which one is climate? What do readers think?
Paul Daniel Ash said:
There are – of course – exceptions to the rule, but I’ve found the correspondence between climate change attitudes and cultural affinities to be very strong in almost every case.
___________
I agree, the correspondence is high (for the extremists on both ends, but less so for the larger group moderates in the middle) and runs the full range from social and religious to political and economic.
Apparently things evolved this way: Weather, then Climate, then Climaterium……
Thus the accuracy of the ensemble forecast is dependent on the accuracy of that single model.
Depends on what you mean by “model.” If you run simulations on the same system (e.g. PIOMAS) with different parameters, different forcings, different initial conditions, I wouldn’t call those the same “model.”
What you seem to be referring to is meta-ensembles: my understanding is that’s been tried and is tricky but promising:
http://journals.ametsoc.org/doi/pdf/10.1175/BAMS-89-3-303
Paul
I have always taken the view that there are very many more important things than AGW to worry about, that can be tackled much more cheaply and with greater overall benefits.
So should we worry about weather-known to knock humanity sideways at times-the climate-surely something more powerful than any of us- or trying to influince things where we can make a difference?
I think you’ll enjoy this article from Bjorn Lomberg in todays Telegraph. I think this would make an interesting thread here together with the opportnity to vote as to where we should conentrate our resources.
http://www.telegraph.co.uk/comment/personal-view/3613517/Save-the-world-ignore-global-warming.html
tonyb
Evan,
No need to guess on my behalf, I’m here to help. As I’ve repeatedly explained, empirical evidence consists of verifiable facts, such as raw temperature data – you know, the kind of raw data that’s so difficult to get, despite FOI requests, and Steve McIntyre’s non-stop efforts, among others. It is clear that mainstream climatologists do not believe in the scientific method.
I’m currently reading Montford’s The Hockey Stick Illusion, and it reinforces the fact that there is almost no empirical evidence in climatology. Everything has been “adjusted” [almost always upward] and getting the original raw data is harder than pulling teeth. Montford pulls no punches, labeling this meddling with the original data “scientific misconduct.”
Also, I have no problem with radiative physics. The problem is again that there is no empirical, testable data showing the extent of any warming due to human CO2 emissions; it is all based on climate models.
If there existed verifiable, testable data showing the impact of increased CO2 on temperature, then there would not be any argument over the climate sensitivity number. The sensitivity number gets ratcheted downward in every new Assessment Report, and based on the planet’s response to the CO2 being emitted both naturally and by human activity, it’s easy to see that CO2 has very little real world effect once the natural warming from the LIA is taken into account.
Excerpts from: Paul Daniel Ash on July 1, 2010 at 7:27 pm
Near as I could tell since I wasn’t here when it was posted, Anthony had you pegged as “PDA” right away. Afterwards you went back to the full name. BTW, your “explanation” doesn’t make sense. Purging the cache, actually the “personal data” like the cookies, leaves a blank comment form. Since “Name” is a required field then automatically hitting “Post Comment” while assuming the fields (which includes “Email” and “Website”) were auto-filled should have caused the post to be rejected. After such purging, I know of no mechanism by which “PDA” would have been auto-filled in that field. Therefore for “PDA” to be there, someone typed it in. Most likely, you typed it in.
If you have an alternate explanation of how “PDA” got in the “Name” field, perhaps some funky weird Windoze auto-fill thing which would seem to indicate you have used “PDA” before, that actually makes sense, or can properly explain how your previous explanation really does make sense, feel free to supply it.
BTW, from your “explanation”:
Now that the “Climate Wars” have escalated to (C)AGW proponents drawing up hit lists, note the last part is not in quotes, anonymity is self-protection from reprisals. Indeed, certain “Doubters” not up to doing Resistance work would be well justified in pulling an Anne Frank for a while.
Another excerpt:
My, aren’t you reaching? You have my first initials and last name, I am in central Pennsylvania “near the river” as I have posted before. I am also in relatively close proximity to at least three (C)AGW-friendly universities, sole caretaker of my mother who is in frail health, as I was to both my parents not long ago, and groups of (C)AGW deluded protesters or even a sole “determined” individual showing up on the property could well be enough stress to kill her. After anywhere from a day to a few years from now that will change, then you will find out the rest.
Smokey, regarding data and proof of CO2 effects.
Does the satellite data of Roy Spencer et al. count? That shows stratospheric cooling and tropospheric warming which cannot be accounted for by any other mechanism than CO2, being on a global scale, and is somewhat in line with what would be expected to be observed.
Jim D,
Yes, Dr Spencer’s data appears to be empirical evidence. However, you shouldn’t overstep by saying that CO2 is the only possible mechanism. That’s the often used argumentum ad ignorantium fallacy here: because we can’t conceive of any other explanation, then the one we have must be the only possible answer. In fact, there are many possible answers.
What we lack is empirical evidence showing the impact of a given amount of CO2 that results in a specific increase in global temperature. We don’t even know the true global temperature, because the raw data has been adjusted and re-adjusted so often [the CET is an exception, which does not agree with the IPCC or Michael Mann’s version of reality]. At this point there is too much we don’t know, which results in guesswork. If we knew the true climate sensitivity number, all this would be moot.
But we are really just guessing at this point, which is why the scientific method is so critical. The alarmist crowd wants us to take their word for it, while the skeptics respond: make your case by using unadjusted empirical evidence. The alarmists’ consternation at having their feet held to the fire of the scientific method is obvious.
Just a note:
I don’t think AGW is correct, but I wish it was, don’t like cold winters, that’s all.
From: Paul Daniel Ash on July 2, 2010 at 10:49 am
“Model” is a glorified name for a program, with all its subprograms and subroutines etc, which can use some data that is not inputted at the start or during the running (the “tweaking” etc). Your argument boils down to saying since different numbers were used then it’s not the same program.
Tricky but promising, yet acceptable for use by Zhang et al? If the method is good enough to stake a professional reputation on it, then it must be considered acceptable at least by those doing so, don’t you think?
BTW, what is this talk of the Global Climate Models being run with different initial conditions? The initial conditions would be the previous weather records, or “climate records” if you insist, with the related info like CO2 concentrations and ocean temps. You have to run a model from a starting point, so you input the starting point. How can you have different “initial conditions” of the same starting point? You can have different runtime parameters to generate an ensemble of predictions, but it’s still the same starting point thus the same initial conditions for all runs in the ensemble.
groweg says:
July 1, 2010 at 6:24 am
….My prediction on that (not computer generated) is that it will destroy our economy and reduce our standard of living to underdeveloped country levels.
_________________________________________________________________________
One of the things everyone seems to forget is underdeveloped countries have a working society and knowledge base that can support their civilization at a lower level of development. Western civilization do not. We do not have the level of expertise or the social structure to keep our society going when our idiotic governments yank the rug out from under us by limiting CO2.
Think about it.
1) The UK is planning to shutting down 30% of its present power capacity by 2030 and replace it with “green energy”
2) the EU has ban the building of heated family homes in the year 2020.
3) And a new report, published today, which features input from 13 universities and 12 research bodies…call for an 80% reduction in livestock numbers
Now add in higher taxes, higher real costs for everything, lower food production because of energy/oil limitations and higher unemployment. Seems to me the politicians have come up with a really really good recipe for anarchy and riots. On the other hand destabilizing a country is the first step in making it vulnerable for take over. I wonder if there is a reason, Maurice Strong, Architect for this madness called CAGW, has moved to China to be an advisor to the Chinese government…..
kadaka asks: “How can you have different “initial conditions” of the same starting point?”
A model usually has much bigger space of initial conditions than “weather records” can provide. For each “weather record” (usually obtained at ground point) one needs to supply the state of entire atmosphere. For each ground “weather record” there could be a large number of slightly different atmospheric “configurations”. Also, “weather records” have finite (and very lousy) accuracy; therefore small deformations of the same “weather record” within the range of data errors would produce different trajectories in a long run, just as Ed Lorenz has originally discovered in his attractor. However, the space of deformations might not be contiguous, and might have “islands” that belong to different attractors, and climate modelers do not have much of an idea that their initial conditions might provide very different outcomes for the same model. The climateprediction.net is an example.
@Enneagram says:
July 1, 2010 at 12:29 pm
“Global warming/climate change has splitted the personalities of many former scientists, who in order to survive in such a competitive environment have surrendered their science replacing it for a creed. This is why some show two or more personalities at the same time, playing the eternal drama of good vs. evil, appearing like Dr.Jekyll and Mr.Hyde.”
So the Alter-Ego (like the alter in a church) of Global warming/climate change dominates the true self of `just a change in the weather` through its feelings of inadequacy of not being able to understand its own true nature and personality (traumatised ego), and hence what it would do next, while needing to appear to be `in the know`, which then provides the need for a social`mask` of respectability…. Enter the alter-ego, whose nature is typically the complete opposite of the self, and totally unable to to see the true self, while being totally at its mercy. This puts increasing pressure on the alter-ego to expand its reasons for its own existance, and its own claims of self importance, to a point of critical mass, where the whole thing has to essentially super nova, as it has become so detached from reality, and is an unbearable strain on relationships.
The harder they come, the harder they fall.
How about using the word `climate` in its traditional sense as a regional norm, and then discuss the weather as it happens (or in advance if you are a good forecaster) and just leave it at that.
Vuk etc says:
July 1, 2010 at 11:04 pm
Vuk etc says:
July 1, 2010 at 11:04 pm
I think both Dr.Jekyll and Mr. Hyde will have to accept this:
http://www.vukcevic.talktalk.net/AMOFz.htm
vukcevic says:
July 2, 2010 at 4:04 am
http://www.vukcevic.talktalk.net/AMOFz.htm
This graph it is astounding, but now you have to relate it to the Sun, as GMF is modulated by it. I would suggest you a real post to decipher all your graphs for us “commoners”.
Enneagram thanks for the note. I will eventually put some units on it, and may write a bit, but at the moment working on something else, just as controversial. Our cat is enneazoe.
Dave F said:
July 1, 2010 at 9:15 am
And aside from that, each coin flip is independent. I would not say that each day’s weather is independent, and you certainly cannot say that climate is independent of weather.
OK, how about tossing a coin where you place the coin on your thumb with the previous result facing upwards, i.e. if the last flip landed heads, start the next flip with heads facing upwards.
This way the result of each flip is not independent of the previous one since the system is strictly deterministic (governed by deterministic newtonian mechanics -unless you want to try and argue that quantum effects are somehow significant).
But it doesn’t make any difference. Now you have a system where each outcome depends on the previous outcome, you cannot predict the outcome of any single coin flip, but you can accurately predict the outcome of a large number of flips.
The logic Steve Goddard used in the article was flawed.
timheyes says:
July 1, 2010 at 3:32 pm
“I think you’ll find that January colder than July in the northern hemisphere due to Earth’s orbit around the Sun and the axial tilt of the Earth WRT the orbital plane.”
So what happened to the orbit and/or axial tilt when, between 1907 and 1949 in The Netherlands; 1947 recorded a January temperature of 17.2 C and July 13 1907 a temp of 13.9? Something is missing? Weather?
@Enneagram says:
July 3, 2010 at 10:31 am
vukcevic says:
July 2, 2010 at 4:04 am
http://www.vukcevic.talktalk.net/AMOFz.htm
This graph it is astounding, but now you have to relate it to the Sun, as GMF is modulated by it. I would suggest you a real post to decipher all your graphs for us “commoners”.
_______________________________________
Thats Gleissberg, call it 32700 days, and check it against the major synodic periods of J/N 4669d, J/U 5045.5d, J/S 7254d, S/N 13102d, S/U 16571d, E/V 583.9219d, E/Mars 779.9356d, E/Ceres 466.7165d etc, so clearly the Sun doing it.
We still need to know about the weather though, the short term range is massive in comparison, maybe 10degC in England in January, the Sun is doing that too.
From: bemused on July 3, 2010 at 7:22 pm
Wrong. For that system to be deterministic you would have to know all initial conditions and all energy transfers that would occur during the flip before the flip took place. To keep it simple, if you can simply crank a set of numbers through the equations and always get the same repeatable result, then a system is deterministic. If you can’t, there is some amount of randomness involved, then the system is stochastic. By invoking a thumb you have invoked a person, thus you would have to know what energies a person would transfer during the launching of the coin, thus you would have to “solve” a person to calculate the result of the flip. Thus the “person” system involved would have to be deterministic. Good luck demonstrating that.
Besides, a “coin flip” is by definition a random event that has only two possible results, with equal probability assigned to each. It is not normally given the messiness of real world coin flips, where differences in the faces will make one side more probable, nor is the on-edge landing considered. Retaining the previous result as an initial condition during a succession of coin flips does not make each flip not independent, as the initial state of the coin does not matter.
Thus the logic of your comparison is flawed.
Wrong. For that system to be deterministic you would have to know all initial conditions and all energy transfers that would occur during the flip before the flip took place.
I assure you that Newtonian mechanics are deterministic. The coin does not magically move itself for no reason. It’s motions are governed entirely by the forces acting upon it.
To keep it simple, if you can simply crank a set of numbers through the equations and always get the same repeatable result, then a system is deterministic. If you can’t, there is some amount of randomness involved
Yes, but an important characteristic of chaotic system is that they are deterministic systems which are observationally indistinguishable from indeterministic systems. If you take a simple non-linear set of equations (for example see Lorenz’s equations), then the outcome is deterministic (i.e. it depends on the initial conditions), but minute differences in those initial conditions (below observational error) will get larger with time and make the two solutions so different that they may as well be random. But crucially, in the Lorenz system you can predict the bounds of the attractor (the climate) even though the internal fluctuations of the system within those bounds are, to all intents and purposes, random.
I don’t want this discussion to degenerate into a metaphysical argument over whether or not humans have free will so, for the sake of argument, why not let a robot arm flip the coin? If the coin is thrown sufficiently high then I am almost certain that the results will not always be repeatable.
If anyone is left reading this thread, here’s one more. I think many people aren’t comfortable with simple analogies (e.g. coin tosses, cups of coffee etc), but maybe some examples using the real atmosphere would be more convincing.
A small parcel of air is made up of billions of molecules travelling in different directions, colliding on seemingly random paths. You can’t predict the exact path of each individual molecule -in fact, you don’t even know the initial conditions of each molecule. So, given that the individual molecules are completely unpredictable how can you possibly predict the statistics of a large sample of those molecules? Steve’s logic would have you believe that it is impossible. How can I possibly predict that if I increase the pressure (while keeping the volume constant) then the temperature (=average kinetic energy of all those molecules) will go up?
Meteorologists can make reasonably accurate short range predictions of mean wind speed and also peak gusts. But gusts are caused by small (10s of metres) eddies in the air which (even if you knew the initial conditions of each individual eddy -which you don’t) are completely unpredictable beyond a few seconds in advance. Steve’s logic would have you believe that because you can’t predict the individual eddies you can’t possibly predict the statistics of many of those eddies over a time period (the peak gust).
A weather forecasting model may be able to accurately predict that in 24 hours there will be convective showers in the North Atlantic behind a cold front. Yet anyone who has tried to make extrapolation nowcasts from radar animations will understand that individual convective showers are unpredictable beyond an hour or two in advance (or slightly more for more organised mesoscale convective systems). Steve’s logic would have you believe that as you can’t predict the individual showers 24 hours in advance, you cannot possibly predict the statistics of many of those showers (e.g. the fact that a number of showers will occur across a defined area and time window).
Hopefully, these examples show that it is possible to make accurate predictions of certain statistics associated with the atmospheric system despite the fact that the individual events that make up those statistics are themselves unpredictable.
If the sun disappears in 2015, I predict that Earth will get colder. I don’t have a clue about the individual weather events that will transpire, but the average of all those events will still give an increasingly cold Earth. Climate prediction is not impossible.
Bemused, you are confusing statistics of a relatively short-term events with individual (unique) trajectory of a long evolving object.
To apply your line of reasoning more coherently, yes, billions of colliding molecules on a billiard with negative curvature gives you a very good ergodic system, such that applying mathematics of big number one can get a clean thermodynamical state of continuous matter (in the limit of thermodynamic equilibrium).
However, when this system is subject to variable forces or is far from equilibrium under some uniform field, the media becomes unstable, and dynamics of individual parcels becomes unpredictable again, as Navier-Stokes equations demonstrate. Again, you need to resort to statistics (of turbulence, invariant measure of a chaotic attractor) to be able to predict dynamics, but only for a time scale that is characteristic for this eddy motion or whatever.
Furthemore, to do so you either need a coherent theory of eddy interactions (which you don’t have due to divergence in Reynolds averaging process), or use experiment to establish these probabilities (eddy viscosity, “universal logarithmic profiles”, Obukhov mixing length, etc.) from observations. To get this statistics for a minute-long process, people run various tanks and wind tunnels for months collecting zillions of spatio-temoral data points, all under very well controlled conditions.
Now, the key point is that when you combine all this stuff into a much bigger system, you have unspecified strange attractor of weather that, according to known facts from nonlinear dynamics, would certainly have long tails of slow-evolving fluctuations. Yes, theoretically one can find invariant measure of this attractor and construct various PDFs out of it, but practically it would require experimental verification. This verification, as usual, should be conducted over time scales that are characteristic for this type of motion.
Typically, even in studies of well-defined strange attractors people would take time series of thousands of characteristic cycles and billions of data points to characterize the attractor statistically. Since we are talking about climate, so far we have only a fraction of climate trajectory. Even to get a single turn around climate attractor one need to get full data for entire atmosphere (and all boundary conditions as well) for at least 100,000 years. Only then you can have some basis for climate prediction.
Re: bemused on July 4, 2010 at 2:57 pm
Several times now I’ve tried composing a reply, it continually gets too wordy. The problem, I have come to realize, is you just don’t understand the simplifications normally done when using deterministic Newtonian mechanics to predict real-world behavior. Predictions are what is being discussed, specifically climate models. You yourself have resorted to simplifications, using the far-simpler system of coin flipping to argue for the accuracy of the vastly more complex global climate models, which themselves are already greatly simplified from the real world climate systems. You have essentially agreed that even coin flipping is too complex for accurate predictions, since when faced with having to know exactly what a person will do while flipping the coin beforehand you have resorted to suggesting using a robot arm, a further simplification.
One may argue that everything is deterministic. True randomness does not exist. In reality, there are too many variables to keep track of. So, one simplifies things down to where predictions can be calculated. In the process you lose precision, a measure of accuracy is lost. Those factors dropped from consideration aggregate into that which is ascribed to chance, to randomness. Yes, in an absolute sense all systems are deterministic. In reality, we settle for “close enough” predictions, normal procedure being the performing of many test runs and invoking statistics to show your method of prediction is indeed close enough.
Thus so far, you have done a good job showing why the global climate models fail. They are a simplification of the actual climate systems to start with. They are “tuned” to provide (reasonably) accurate hindcasts, but have not been show to provide “close enough” predictions. Due to simplifications, they ascribe many things to randomness thus a stochastic nature is presented of what ideally is a deterministic solution.
Heck, with the repeated iterations even the rounding off of the numbers can grow into a significant source of inaccuracy.
Now if there is this much difficulty in applying deterministic Newtonian mechanics to accurately predicting something as relatively simple as coin flipping, how can you think that vastly complex global climate models can accurately predict conditions for decades and even centuries past the starting point?